Nano BTE

[jk262626]

Hi Nakib,

I have a question on nano BTE parameters.
I want to calculate thin-film thermal conductivity for GaN in both in-plane and cross-plane (thickness direction).
Thickness range I want to study is from 1nm to 300nm.
In this case, can I set the parameters like this? I don't think I fully understood 'limit' parameter.

nsys = 1
tag = ‘tf’
limit = 300
taxis = 1 0 0
naxis = 0 0 1

[nakib]

Dear Jiyoung,

Thanks for your question. In a first glance, your input looks correct to me. I am guessing you are looking to compute the thin-film transport in the x-hat direction which is why you chose taxis = 1 0 0, naxis = 0 0 1. For the limit (that sets the characteristic size of the particular type of nanostructure), you chose 300 nm. If you want to do 4 different sizes, you will have to set limit = 1.0 100.0 200.0 300.0, for example.

I am tagging @mrm24 here since he wrote the nanostructures stuff. Perhaps he can confirm that your input is correct.

Best,
Nakib

[mrm24]

Dear Jiyoung,

Your input is correct but as Nakib commented it is only doing a 300 nm thick thin-film. Imagine you want to compute it for a linear range between 1 and 300 nm, then the nano block would look like:

&nano
nsys = 10
tag(1:10) = "tf" "tf" "tf" "tf" "tf" "tf" "tf" "tf" "tf" "tf"
limit(1:10) = 1.0 34.222 67.444 100.667 133.889 167.111 200.333 233.556 266.778 300.0
taxis(1,:) = 1 0 0
taxis(2,:) = 1 0 0
taxis(3,:) = 1 0 0
taxis(4,:) = 1 0 0
taxis(5,:) = 1 0 0
taxis(6,:) = 1 0 0
taxis(7,:) = 1 0 0
taxis(8,:) = 1 0 0
taxis(9,:) = 1 0 0
taxis(10,:) = 1 0 0
naxis(1,:) = 0 0 1
naxis(2,:) = 0 0 1
naxis(3,:) = 0 0 1
naxis(4,:) = 0 0 1
naxis(5,:) = 0 0 1
naxis(6,:) = 0 0 1
naxis(7,:) = 0 0 1
naxis(8,:) = 0 0 1
naxis(9,:) = 0 0 1
naxis(10,:) = 0 0 1
/
So you need to provide the kind, the limiting length (tf thickness in your case), the transport axis, and the normal axis (this last one is not needed for the nanowires), for each nanostructure.

Please note that you will only get the in-plane conductivity.

Moreover, keep in mind that the implemented approach ignores any quantization or confinement effect over phonons for extremely "thin" thin films.

Best,

Martí

[nakib]

Thanks, @mrm24 for the clarification.

@jk262626, you also have the option of using a simple model of phonon-thin-film scattering. (These are not fully documented yet...sorry.) For this, set thinfilm_height = 0.100e-3 !h = 300nm and thinfilm_normal = 'z' in the crystal_info namelist. You can also specify the specularity factor there by setting specfac. It is zero, by default. Finally, set phthinfilm = .t. in the numerics namelist. There is a related parameter called phthinfilm_ballistic, which is by default true. This employs the large Knudsen number limit of the Fuchs-Sondheimer (FS) theory. If you set it to false, the more general theory will be used. The general FS theory has not yet been thoroughly tested. So use with caution and do your own tests if you want to use that!

Best,
Nakib

[jk262626]

@mrm24 Thanks a lot for the clarification!
@nakib Thank you so much for the additional information! Regarding the current calculation that's documented—how is the thickness size effect being accounted for (I mean,not the simple model using FS theory)? Is there any documentation that explains the theoretical basis, like in the first elphbolt paper? Or should I look through the code to figure it out for now?

[nakib]

I think the nanostructures stuff is explained in the following paper by @mrm24 and his coworkers: https://www.sciencedirect.com/science/article/pii/S0017931024002163

Best,
Nakib

[jk262626]

Thank you for the information Nakib!

@mrm24 I submitted calculation with the input file as you suggested, but got an error message in phonon setting step like this:

___________________________________________________________________________
______________________________________________________Setting up phonons...
Precomputing q-independent quantities related to dynamical matrix...
Calculating phonon FBZ quantities...
Calculating IBZ and IBZ -> FBZ mappings...
 Number of FBZ wave vectors =     125000
 Number IBZ wave vectors =       6084
Calculating phonon mesh triangles...
Reading ShengBTE format third order force constants...
 Number triplets read in =       1444
..............
| Timing info: Phonons  0.12263376E+01 hr
..............
At line 135 of file ././src/nano.f90 (unit = 1, file = 'input.nml')
Fortran runtime error: End of file

Error termination. Backtrace:
At line 135 of file ././src/nano.f90 (unit = 1, file = 'input.nml')
Fortran runtime error: End of file

Let me explain my input file structure. I included only nanoBTE related part.
Do you think there is any problem with this?

&allocations
/

&crystal_info
/

&electrons
/

&numerics
      solve_bulk = .False.
      solve_nano = .True.
/

&wannier
/

&nano
nsys = 2
tag(1:2) = "tf" "tf" 
limit(1:2) = 2.0 300.0
taxis(1,:) = 1 0 0
taxis(2,:) = 1 0 0
naxis(1,:) = 0 0 1
naxis(2,:) = 0 0 1
/

[nakib]

Dear Jiyoung,

I was able to run a nanobte without any problem with your &nano namelist information. Not sure why you are getting that error.

Best regards,
Nakib

[jk262626]

I reinstalled elphbolt and the problem was resolved. Thank you!